{"resourceId":"maastricht-wibral-ai-exercises-to-protecting-learning-20261009","versions":[{"version":"external-d1a567ad2fb026d4aba5f6a58d65255512f70f42cfbe8ad2d6c8d39f320c266c","resource":{"id":"maastricht-wibral-ai-exercises-to-protecting-learning-20261009","title":"Maastricht University: instructor's grant-funded ChatGPT exercises shift to protecting unaided learning, including no-computer first-year tutorials","organization":"Maastricht University School of Business and Economics (Matthias Wibral; EDLAB Education Innovation Grant)","sector":"Higher education student success / first-year course design and generative AI","geography":"Netherlands (single compulsory first-year course, about 750-800 students)","publishedAt":"Maastricht University news, 9 October 2026","publicationDate":"2026-10-09","eventDate":null,"sourceName":"Maastricht University news","sourceLabel":"University news feature (practitioner case narrative; no measured outcomes)","sourceUrl":"https://www.maastrichtuniversity.nl/news/teaching-students-use-ai-protecting-learning","evidenceClass":"independent-reporting","outcomeClass":"emerging","topics":["knowledge-work","operating-model","accessibility-workforce"],"finding":"An associate professor teaching a compulsory first-year introductory economics course of about 750-800 International Business students used an EDLAB Education Innovation Grant to buy out teaching time and design exercises encouraging more critical ChatGPT use: AI-assisted LaTeX graph coding, optional reflection on AI concerns, and students building course chatbots that peers tried to make answer incorrectly. Students became fluent with generative AI faster than expected, and improving models made the error-hunting exercise harder. His focus shifted from teaching AI use to deciding when students must work without it; he now runs first-year tutorials with no computers, and added an alternative assignment for students who object to using AI on ethical grounds.","sledRelevance":"Published 9 October, this is a current instructor-level account of a course-design pivot many campuses face: AI-literacy exercises can age quickly, and the student-success risk moves to preserving independent practice in gateway courses.","evidence":"Practitioner narrative from one course; no grades, engagement data or evaluation reported. Course size about 750-800 students.","architectureImplications":"No platform change; interpretation: course-built chatbots and error-hunting tasks depend on model behaviour and need periodic redesign as models improve.","governanceImplications":"Shows course-level AI policy (no-computer tutorials, AI-optional alternative) being set by instructors; interpretation: institutions need guidance that allows AI-free practice and ethical opt-outs.","securityPrivacyImplications":"Privacy is mentioned only as an optional reflection topic; student-built chatbots would need data-handling rules if course materials or student data are uploaded.","caveats":"Single instructor's self-report in an institutional feature; no outcome measurement; European business-school context; transferability to U.S. gateway courses is untested.","streamIds":["student-success"],"roles":{"sales":"The framing point is that students' AI fluency now outpaces many AI-literacy exercises. A Maastricht instructor teaching about 750-800 first-year students found students know ChatGPT too well but struggle without it, and moved tutorials to no-computer settings. Use this in conversations with teaching centers about course redesign support, not tools. Keep claims bounded: one instructor's narrative, no measured outcomes, and a European business-school context.","engineering":"For engineering, the framing point is that AI-dependent exercises decay as models improve. Peer error-hunting on course chatbots stopped working well once models hallucinated less. Any campus AI-literacy tooling should be versioned and re-tested each term against current models, and courses need a supported AI-free mode for assessments. Validate by checking whether planned exercises still elicit model errors and whether AI-free practice can be scheduled and proctored at scale.","delivery":"Delivery's framing point is resourcing redesign rather than writing policy. Fund small faculty buyouts for gateway-course redesign, provide an AI-free practice option and an ethical opt-out alternative, and connect instructors through cross-faculty sessions. Pilot in a few large first-year courses and measure unaided performance. Treat the Maastricht account as a design prompt, since it reports no outcomes and comes from a single course."},"roleDetails":{"sales":{"customerProblem":"Gateway-course instructors see students relying on AI and lack time to redesign assessments that preserve independent practice.","stakeholders":["Vice provost for teaching and learning","Teaching and learning center","Gateway-course department chairs","First-year experience office"],"discoveryQuestions":["Which large first-year courses report students struggling without AI?","Is there funded time for instructors to redesign assessments?","Do courses offer AI-free practice and ethical opt-out options?"],"valueHypothesis":"Funded redesign of gateway courses could protect independent practice; benefit is unmeasured in this source.","potentialEngagement":"A bounded course-redesign support pilot for two or three large first-year courses.","unsupportedClaims":["That no-computer tutorials improve grades","That AI-literacy exercises are ineffective in general","Any retention or performance effect"]},"engineering":{"fit":"Campuses running AI-literacy exercises or course chatbots in large first-year courses.","architectureIntegration":"Versioned course chatbot configurations and assignment templates, re-tested each term; LMS support for AI-free proctored practice.","prerequisites":["Approved campus AI tool","Assessment scheduling for AI-free sessions","Instructor time"],"constraints":["Exercises depend on current model weaknesses","Single-course European evidence","No outcome data"],"security":"Restrict what course materials and student data are loaded into student-built chatbots.","proofOfValue":["Re-test planned AI exercises against current models before term","Compare unaided quiz performance in sections with and without AI-free practice","Track opt-out alternative uptake"]},"delivery":{"work":"Run a funded redesign of selected gateway courses adding AI-free practice and AI-optional alternatives.","dependencies":["Innovation grant funding","Teaching center designers","Department approval"],"ownership":"Teaching and learning center with gateway-course instructors.","skillsAdoption":"Instructors learn to judge when AI adds to learning and to design tasks requiring students to check AI output.","governanceCheckpoints":["Course AI policy review","Accessibility review of no-computer settings","End-of-term evaluation review"],"acceptanceCriteria":["Each pilot course publishes when AI is and is not allowed","Unaided performance measured in pilot sections","Opt-out alternative available in all pilot sections"],"risks":["No-computer settings disadvantage students needing assistive technology","Exercises age within a term","Overgeneralizing from one instructor's experience"]}},"retrievedAt":"2026-10-11T03:16:01Z","enrichedAt":"2026-10-11T03:18:59Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Faculty need protected time (here grant-funded buyout) to redesign assessments as AI capabilities change.","procurementImplications":"Limited relevance; suggests not over-investing in fixed AI-literacy exercises tied to a specific model's weaknesses.","operatingModelImplications":"Small innovation grants plus cross-faculty education days gave the instructor time and peer input; interpretation: a repeatable model for course redesign.","sourceVerification":{"openedUrl":"https://www.maastrichtuniversity.nl/news/teaching-students-use-ai-protecting-learning","referenceExcerpt":"Now I think my problem is more they know too well how to work with ChatGPT. They don’t know how to work without ChatGPT anymore.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}